56 ALASKAN GLACIER STUDIES been impossible in 1906 and even travel on foot could be undertaken only by the most difficult ice work. We traversed the eastern margin of Malaspina Glacier from the point of emergence of Marvine Glacier to the mouth of the Kwik River, and along this entire distance, excepting only where dominated by Hayden Glacier, the ice was greatly broken. Prior to 1906 this land margin had been covered with ablation moraine bearing a forest of cottonwood and alder and some spruce along the margin of the lower Kwik valley, and this margin was so moderately sloping that it could easily be ascended at any point; in 1906, however, the margin was for the most part changed to a steep and broken cliff down which the ablation moraine was constantly sliding, while in the forest-covered area trees and bushes were overturned and were falling into crevasses or crashing down the newly steepened ice face (PI. XV, A). The cliff was crevassed, often, into impassable condition, and the broken ice blocks protruded through the ablation moraine, giving the appearance of a frost-riven rock cliff, heightened by the earthy stain that the angular ice blocks bore, and by the trees that still stood among them. The ablation moraine, which had covered the ice for a depth of from 2 to 15 feet, and had so protected it from melting that sufficient stability was secured for the growth of trees to maturity, was now sliding into crevasses and down the ice cliff, often in mud flows. Abundant water (PI. XVI, A) was now supplied by the melting of the ice, newly-exposed to the air and rains, and small and large streams, heavily charged with morainic de'bris were issuing from all portions of the ice margin, where hitherto only trickling streams of clear water slowly oozed from the moraine-covered glacier. That the ice was being broken and crowded forward during July and August, 1906 was evident as we passed along its base; and that the entire breakage was a result of changes during that year was indicated by the fact that among all the trees found overturned, not one was seen that had not fully developed foliage (PL XV, B). This was the last season for many trees, for in August some were dead and others had yellow leaves, as if dying. This evidence is in harmony with several other facts which indicate that this advance was confined mainly to 1906. These facts are as follows: (1) Our observations of 1905, though not very full, indicate that the advance had barely begun; (2) other glaciers in the region underwent similar absolute transformation in the brief period of nine months; (8) the slight amount of ablation on the broken surface of the Marvine lobe, already mentioned; (4) the closing of the subglacial tunnel by which the waters of the Blossom Island region had formerly escaped, which occurred between the autumn of 1905 and the middle of July, 1906, and which gave rise to the formation of a temporary expansion of the lake which lies in the Blossom Island depression.1 Thus a long period of stagnation of sufficient duration to permit the growth of a forest with trees a half century old, and in other portions of the glacier a period of such moderate motion as to permit the development of a smooth ice surface over which travel was' easy in all directions, was abruptly ended by such a sudden and rapid forward motion that even the most stagnant portions of the ice margin were broken. Whether the ice was thickened by the advance could not be determined, for we had no exact basis for measurement, but, in view of the fact that other glaciers which advanced were thickened, it is probable that this one was also. There was certainly some advance along the margin, for only by this could the ice cliff be steepened and broken, and ice blocks caused to > Tarr, E. S., Professional Paper, U. S. Geol. Survey, 1909, p. 88.